Smart Document Engine quickly and precisely classifies questionnaires and various forms in the stream of images and performs automatic recognition of typed and hand-printed details, as well as graphic fields. With the help of innovative computer vision and AI algorithms, Smart Document Engine automatically detects a document in an image, corrects distortions and skewness, automatically determines a document type, finds text and graphic details, and recognizes them using the innovative GreenOCR® optical text recognition technology. Smart Document Engine software recognizes documents precisely and quickly both from images captured from flatbed or sheetfed scanners, and from photos captured from document cameras or regular smartphones and tablets.
Smart Document Engine is a versatile system for recognition of flexible and fixed document forms, it provides automated data entry and verification of various forms and questionnaires with high speed and precision. The Smart Document Engine solution can be applied for automatic data entry of documents both in offices and service points, as well as for remote service – in such use cases as data recognition of consents to personal data processing, proxy forms, and other onboarding forms. The Smart Document Engine form classification module allows you to implement completeness verification when entering and registering a package of documents. Automatic recognition of complexly structured questionnaires and forms allows you to simplify your onboarding process — insurance, banking, financial, or legal services; personal data entry when passing medical tests; exam and test data processing; traffic accident data entry, public-service submission application, and many other processes.
Thanks to deep algorithmic and code optimization, the Smart Document Engine SDK includes the feature to scan and recognize questionnaires and forms with fixed layouts in any application for mobile phones and tablets, office applications, an integration bus, and data entry centers. It’s worth noting that the SDK is not resource-greedy, questionnaire scanning and recognition even on a regular smartphone takes about 2 seconds. Thus, a smartphone turns into a high-performance document scanner. Camera requirements are not high as well — a regular 8-megapixel autofocus camera is enough for high-quality recognition. The system performs automatic OCR of typed fields, automatic ICR of hand-printed fields, automatic OMR for labels and automatic recognition of 1D and 2D barcodes. A tabular data recognition module, as well as a module for extracting and verifying graphic details (such as stamps and signatures) have been implemented in the Smart Document Engine system. To recognize a document with a fixed layout, the system implements a zero-shot learning approach, i.e., the system does not retrain models on large amounts of data, a few examples are enough to automate data entry processes.
The Smart Document Engine system is an independent on-premises solution and does not transfer images to any third-party servers for processing and recognition, thus, sensitive data transfer to third parties is excluded. Consequently, Smart Document Engine is the most secure solution for both business and its customers. The Smart Document Engine SDK does NOT transfer data and images for processing to Smart Engines or to any third-party services, does NOTsave data or images (the processing is carried out in the device’s local RAM) and does NOT require internet access.The Smart Document Engine SDK is GDPR, CCPA, HIPAA-compliant.
Smart Document Engine is delivered as an autonomous SDK (software development kit) to integrate it into the customer’s infrastructure and applications. The SDK includes OCR libraries compiled and optimized for a set of target platforms and operating systems, API documentation (application programming interface) and integration examples for various programming languages. The scanning and recognition API for questionnaires and forms allows you to integrate functionality into programs and applications in the programming languages C, C++, C#, Java, Python, PHP, Objective C and Swift. Optionally, the SDK can be bundled with the Smart ID Engine and Smart Code Engine systems to scan and recognize ID documents and codified objects using a unified software interface.
Thanks to simple API, the Smart Document Engine SDK can turn iOS or Android OS phones into a scanner of questionnaires, consent and proxy forms, and other documents with fixed and flexible layouts. Our proprietary AI and computer vision algorithms provide high-speed (about 2 seconds per recognition of a full-size photo on iPhone XR) and autonomous document recognition in a wide variety of languages.
Deep software and hardware optimization of the code and algorithmic recognition principles for various processor architectures allows Smart Document Engine to autonomously recognize questionnaires and forms on any personal computer (including nettops and all-in-one PCs), and turn them into high-performance workstations for entering and processing document images captured from various image capture devices (flatbed and sheetfed scanners, specialized scanners and document cameras). A wide range of software interfaces provides the possibility of native integration into most information systems.
Thanks to thread safety and efficient resources management, the Smart Document Engine SDK allows you to implement scalable AI platforms for image processing of questionnaires, forms, and other documents with fixed and flexible layouts. Unique algorithms for automatic document classification and localization in images (scans and photos) allow you to automatically process heterogeneous streams of input images. The recognition performance of complex data structured documents on modern computers (with an AMD Threadripper 3970X processor) is about 12 full-size photos per second, which allows our customers to create highly secure services with WebAPI and integrate the SDK into existing data exchange buses.
NEC de Colombia
Travizory Border Security selected Smart Engines product for identification documents scanning
Please fill out the form to get more information about the products,
pricing and trial SDK for Android, iOS, Linux, Windows.